(DI-1908) Alert Definition

Owned by SNP
Apr 05, 2019
5 min read

Definition of alerts

There are three options how to define an new alert.

1. The simple one is to use the context menu of a KPI and choose to create an alert.

2. The second option is to define alerts in monitoring customizing by selecting KPI from KPI defintions and choosing the option Assign alerts. There you can create new entry for an alert definition or edit an existing one. You can create only an alert of the threshold kind. For each newly created alert you must define at least one threshold value.

3. The third option is to define alerts in monitoring customizing by selecting Alerts and defining the bellow mentioned fields.



Alert view

This definition consists of:

  • Alert name - a unique, case sensitive technical name of an alert.
    Please note that you can create or modify only field names (in this case the alert name) that start with Z*, Y* (or X* for generated Alert name). The standard customizing (field entries except the ones starting with Z*, Y*, X*) can not be changed or added.
  • Alert description – the description is used as the label in outputs.
  • KPI name and description – name of KPI for which an alert is created.
  • Alert kind - defines the kind of the alert (Threshold or Anomaly) .
  • Alert type – describes the type of alert and determines the icon to be displayed if the alert occurs. Possible values are:
    • Success – 
    • Warning – 
    • Issues – 
    • Error - 
  • Period time – the frequency that specifies how often the alert is rechecked.
  • Changelog - displays the user name and time, when the alert was created and changed.

For every alert created, it is necessary to define associations to the KPI and to the Profile. This can be done in the Threshold setup category of this screen, according to the type of alert. 

Threshold setup

The following fields should be filled:

  • Monitoring profile – alert will be checked only for this profile under running Collector job.
  • KPI name – alert will be checked only for the value of this KPI.
  • Comparison operator – determines in which case a KPI value causes an alert to occur.

Comparison operator

  • Threshold – determines the limit value for a KPI. This alert will occur if the KPI exceeds its defined threshold of associated alert.
  • Unit – determines the unit of measurement to be used, i.e. %, MB etc.
  • Duration start – how long can a KPI value exceed the threshold (in grey area) before an alert will occur (starting delay for alert).
  • Duration end – how long a KPI value has to be outside of the threshold area (in grey area) before an alert will be stopped (ending delay for alert).



Treshold setup

Anomaly view

This view serves only to display existing anomaly alert settings. By this view is not possible to create a new settings for anomaly alert. In case you want to create a new settings or change the existing one, you need to use context menu for KPI in Monitor cockpit. 

  • System ID - alert will be checked only for this system under running Collector job. In case that this field is empty, alert will be checked for all systems. 
  • Monitoring profile – alert will be checked only for this profile under running Collector job. In case that this field is empty, alert will be checked for all profiles under running Collector jobs. 
  • Parameter name - define a name of setting which is used in anomaly detection. Each anomaly alert consist of the following settings:
    • Detection type - specify a method for anomaly detection
    • High percentile - define a value of high percentile which is used in anomaly detection calculation method 
    • Low percentile  - define a value of low percentile which is used in anomaly detection calculation method 
    • Moving window - define a value of how many KPI values in the history should be included in anomaly detection calculation method 
    • Percentile type - define which KPI values will be excluded from anomaly detection calculation method 
      • L (Low) - only those KPI values will be excluded which are less than a value of low percentile in analysed KPI values 
      • H (High) - only those KPI values will be excluded which are higher than a value of high percentile in analysed KPI values 
      • M (Low + High) - substraction of high and low percentile values. In this case is used IQR method for anomaly detection. 
    • Status - specified a status of settings for anomaly detection
      • A (Active) - anomaly detection is enabled when Collector job is running 
      • W (Waiting) - anomaly detection is enabled when Collector job is running, but there are not enough KPI values collected yet. Number of collected KPI values is less than a size of moving window. 
      • C (Closed) - anomaly detection is disabled
      • S (Simulation) - these settings were created when was executed the simulation of anomaly detection on historical KPI values. 
  • Parameter value - define a value which is used in anomaly detection calculation method 

Anomaly view

Notification 

Each alert occurrence can be notified by email or by message in SAP Business Workplace. That can be done in the Alert notification category of this screen. The following fields should be filled:

  • Monitoring profile
  • Recipient address – the email address or username from SAP system can be specified.
  • Type of notification – determines the notification method. Possible values are:
    • E-mail – notification will be sent to the email specified in "Recipient address" field.
    • SAP Office message – notification will be sent to the SAP Business Workplace for user who is specified in "Recipient address" field.
    • Private distribution list – notification will be sent to all addresses defined in SAP Private Distribution List (transaction SO15 or button "Distribution list" in Insights toolbar).
    • Shared distribution list– notification will be sent to all addresses defined in SAP Private Distribution List (transaction SO15 or button "Distribution list" in Insights toolbar).
  • BAdI ID – [optional] ID, which will be used for filtering of BAdI call. You can create your own BAdI for notification using BAdI implementation /DVD/MON_BADI_DEF_EVENT. Each BAdI implementation has to define the filter, which is also used for determining explicit call of the BAdI for processing of notification. Default BAdI implementation used for notification has initial BAdI ID.
  • Alert notification – determines the period for sending of notifications. Possible values are:
    • At first occurrence – notify when alert occurs first time
    • At next occurrence – notify each occurrence of alert
    • At end of alert – notify when alert is completed



Alert notification